[Wien] phonon_lapw
Peter Blaha
pblaha at zeus.theochem.tuwien.ac.at
Tue Apr 6 16:51:34 CEST 2004
2 comments for that:
a) do you use the simple run_phonon script ? Of course this script runs
sequentially all "cases" (case_1,2,3,....) one after the other. If you have
enough processors you may run all "cases" at once (either modifying the
script (put eg. qsub into it and do the save later,...) or executing the
jobs manually on different processors,....
b) If you then still have more processors than (cases*k-points) (you are
very lucky!!), set the scratch directory differently for the corresponding
job: either use ./ for SCRATCH or e.g.
setenv SCRATCH /tmp/pblaha/D5
run_lapw -fc 0.x ...
and
setenv SCRATCH /tmp/pblaha/D10 for the other job. (of course you
should create these directories first).
(You can always redefine SCRATCH and it is not necessary that one uses
SCRATCH as defined in .cshrc)
Regards
> I recently started to work on phonons with wien2k combined with the
> Phonon program of Parlinski. I use k-point parallelisation. My cells are
> so large that I have less k-points in the irreducible Brillouin zone as
> I have processors. Therefore I started calculations with different
> amplitude of the displacement. For each amplidude the init_phonon_lapw
> script generates a series of calculations which are called case_1 ...
> case_x. If you start now the calculations for both series, one or both
> calculation will be not correct or will crash, because the vector files
> in the scratch directory have the same name. It would be saver to extent
> the name of case_1, case_2 ... with the name of the associated .d45
> file. (e.g. case_1 to D5_case_1, when the d45 file was called D5.d45).
> Maybe the authors can think of changing this in init_phonon_lapw and
> analyse_phonon_lapw.
P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671 FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------
More information about the Wien
mailing list